Taxonomic Revision Of Cistanche Deserticola Ma And Cistanche Salsa (C. A. Mey.) Beck Based On Specimen, Morphological, And Molecular Characteristics
Jan 20, 2026
ABSTRACT: OBJECTIVE To clarify the taxonomic status of Cistanche deserticola Ma (C. deserticola) and Cistanche salsa (C. A. Mey.) Beck (C. salsa), establish a reliable identification system, and support precise traceability within the industry. METHODS Based on a review of botanical literature and specimen verification, morphological comparisons were conducted using online herbarium databases. Field investigations were carried out to examine external morphology (including scale leaves and floral characteristics) and cross-sectional structural variation among plants associated with different host species, in order to assess their taxonomic significance. For molecular phylogenetic analysis, plastid matK and mitochondrial matR gene sequences were aligned using MAFFT, and maximum likelihood trees were constructed with IQ-tree. Arabidopsis thaliana and Nicotiana tabacum were set as outgroups for matK, while Arabidopsis thaliana and Solanum lycopersicum were used for matR. Branch support was evaluated with 1,000 bootstrap replicates. Integrating morphological and molecular systematic evidence, the taxonomic circumscription and interspecific relationship between the two species were clarified. RESULTS Historical literature review indicated that the records and medicinal use of C. salsa predated those of C. deserticola, and that the two taxa were long conflated in practice before gradually being distinguished as "C. deserticola" and "C. salsa." Morphological comparisons revealed only minor differences and a high degree of overlap in characters such as bracts, calyx, and corolla tube morphology; both taxa displayed multi-host parasitism (e.g., hosts in genera such as Haloxylon and Suaeda), resulting in substantial morphological convergence. Phylogenetic reconstructions from plastid, mitochondrial, and nuclear markers recovered C. deserticola and C. salsa as well-supported, closely related monophyletic lineages. CONCLUSION C. deserticola and C. salsa constitute a closely related species complex. The integrative taxonomic evidence provided here supports clarification of their taxonomic status and lays a foundation for developing reliable identification methods and precise, species-level traceability for multi-origin medicinal materials.
KEYWORDS: Cistanche deserticola Ma; Cistanche salsa (C. A. Mey.) Beck; literature research; morphological characteristics; molecular sequence
Cistanche deserticola Ma
The genus Cistanche (Orobanchaceae) is an important group of plant resources in China, with about 22 species distributed worldwide and approximately four species in China, mainly occurring in Inner Mongolia, Ningxia, Gansu, and Xinjiang. Medicinal records of Cistanche date back to the Divine Farmer's Classic of Materia Medica (Shennong Bencao Jing), with subsequent mentions in classical works such as Wu Pu's Materia Medica, Collected Annotations on the Materia Medica, and Compendium of Materia Medica. The 2025 edition of the Chinese Pharmacopoeia records the medicinal part of Cistanche as the fleshy stem, which tonifies kidney yang, nourishes essence and blood, and moistens the intestines to relieve constipation; it is used to treat kidney yang deficiency, deficiency of essence and blood, impotence and infertility, soreness and weakness of the waist and knees, debility of tendons and bones, and intestinal dryness with constipation. Owing to its high medicinal and nutritional value, Cistanche is known as the "ginseng of the desert." It contains a variety of chemical constituents, including nucleotides and their derivatives, amino acids and their derivatives, saccharides, alkaloids, terpenes, hormones, flavonoids, and phenylethanoid glycosides, with phenylethanoid glycosides being the main active components. Pharmacological studies indicate that Cistanche enhances the activity of free radical scavenging enzymes, exerting antioxidant effects, can improve sexual dysfunction caused by kidney deficiency, and exhibits glucocorticoid-like properties that regulate immune function. Other studies report that Cistanche promotes memory and ameliorates memory impairment induced by chemotherapeutic agents.
In China, taxonomic ambiguities within Cistanche have long perplexed botanists and pharmacognosists. Professor Ma Yuquan's "Preliminary Study on Cistanche Plants in Inner Mongolia" formally adopted "Rou Cong Rong" as the Chinese generic name for Cistanche and recognized four species in Inner Mongolia: C. salsa, C. sinensis Beck, C. ambigua (Bunge) Beck, and a new species, C. deserticola. In 1992, the Flora of Desert Plants of China listed five species of Cistanche in Chinese desert regions: C. sinensis, C. lanzhouensis Z. Y. Zhang, C. tubulosa (Schenk) Wight, C. deserticola, and C. salsa. In the monograph "Cistanche and Its Artificial Cultivation," Professor Liu Guojun recognized eight species domestically, adding C. ningxiaensis D. Z. Ma & J. A. Duan and C. fissa (C. A. Mey.) Beck. Currently, four species are generally accepted domestically: C. deserticola, C. salsa, C. tubulosa, and C. sinensis, among which C. deserticola and C. tubulosa have been continuously included in the 2005, 2010, 2015, 2020, and 2025 editions of the Chinese Pharmacopoeia. However, as a medicinal plant, C. salsa was only included in the 1963 edition of the Chinese Pharmacopoeia and in local standards of Inner Mongolia, Gansu, Ningxia, and Xinjiang, and its taxonomic status remains controversial.
Clarifying the taxonomic status of C. deserticola and C. salsa has thus become urgent. Our project's preliminary research showed that, compared with other congeneric species, C. deserticola and C. salsa exhibit high homogeneity in distribution patterns, ecological niches, chemical composition, morphological characteristics, and genetic background. At the macro level, MaxEnt modeling indicated that their potential ecological suitability zones in China highly overlap, showing significant ecological convergence. Fingerprint analysis revealed that the chromatographic fingerprints of C. salsa and C. deserticola are similar, with total phenylethanoid glycoside content highest in C. salsa-particularly echinacoside and verbascoside-demonstrating a quality advantage of C. salsa. However, existing evidence lacks systematic integration. There is an urgent need for more comprehensive, in-depth, multidisciplinary evidence that integrates key biological data based on extensive sampling and verification of type specimens. We therefore propose to conduct varietal research and taxonomic revision of C. deserticola and C. salsa from the perspectives of philology, morphology, and genetics to resolve scientific issues in the Cistanche industry, including confusion over varietal origins, inconsistent quality, and resource endangerment.
1 Materials and Methods
1.1 Materials
Field surveys were conducted to collect Cistanche plants parasitizing different hosts, including Haloxylon ammodendron, Krascheninnikovia ceratoides, and Kalidium foliatum. We investigated scale leaves, floral and other external morphological traits, as well as cross-sectional differences, to explore their taxonomic significance. Experimental materials of C. deserticola, C. salsa, C. tubulosa, and C. sinensis were collected from Northwest China and the coastal region of Shandong, and authenticated as genuine by Professor Lin Yulin of the Chinese Academy of Medical Sciences. The sample list is provided in Table 1.
Supportive Service Of Wecistanche-The largest cistanche exporter in China:
Email:wallence.suen@wecistanche.com
1.2 Philological (Literature) Research
We carried out philological studies, including textual research in classical materia medica, and consulted and compiled Chinese and English sources such as the Flora of China. We examined more than a thousand specimens from multiple major herbaria at home and abroad, including type, voucher, and photographic specimens (herbarium codes: ASSAM, B, BM, BSA, C, CAL, CALI, CDBI, E, GACP, GH, HGAS, F, K, KATH, KUN, KYO, L, LD, LE, MICH, MO, NY, P, PBL, PE, PUN, PYU, S [S-P], SABG, SING, SZ, TAI, TI, U, UPS, US, W, and WU), and accessed online databases such as JSTOR Global Plants (http://plants.jstor.org), the New York Botanical Garden's Virtual Herbarium (http://sweetgum.nybg.org/science/ih/), the National Specimen Information Infrastructure (NSII), and the China Digital Herbarium.
1.3 Molecular Systematics
We used MAFFT to align plastid matK genome sequences of 34 Orobanchaceae species, including Cistanche, for phylogenetic analysis. Arabidopsis thaliana and Nicotiana tabacum were used as outgroups, and maximum likelihood trees were constructed using IQ-tree. Similarly, we used MAFFT to align mitochondrial matR genome sequences of 11 Lamiales species, including Cistanche, for phylogenetic analysis. Arabidopsis thaliana and Solanum lycopersicum were used as outgroups, and maximum likelihood trees were built with IQ-tree. Bootstrap support was assessed with 1,000 replicates, and bootstrap values for each branch are shown next to the corresponding nodes.
2 Results
2.1 Textual Research in Materia Medica and Historical Literature
We reviewed and synthesized a large corpus of materia medica texts and related literature, summarizing a century-long history of Cistanche research in China and a millennium-long medicinal history of Cistanche. The century-long development of Cistanche taxonomy in China (Figure 1) can be divided into two stages. Stage 1: Establishment. In 1930, Beck reported three Cistanche species in China, namely C. salsa, C. sinensis, and C. mongolica G. Beck. Stage 2: Development. Since Ma Yuquan's publication of C. deserticola in 1960, Chinese scholars have successively proposed new species and varieties; in 2000, Zhang Zhiyun merged C. ambigua into C. salsa. Studies have also shown rich infraspecific variation within C. salsa and C. deserticola, such as white-flowered C. salsa and "youcongrong" (oily forms). A review of successive editions of the Chinese Pharmacopoeia shows that the recorded original plants of Cistanche have changed over time: the 1963 edition included C. salsa; from 1977 to 2000 only C. deserticola was included; starting in 2005, both C. deserticola and C. tubulosa were included. Notably, although not continuously included, C. salsa has consistently circulated in the market under the commodity name "Rou Cong Rong," and in Mongolian medicine under the names "Gaoyao" (or "Zhagan").
The millennium-long medicinal history of Cistanche (Figure 1) shows that Cistanche has a long-standing medicinal use (Figure 2, Table 2). The earliest record appears in the Eastern Han Divine Farmer's Classic of Materia Medica, stating: "Rou Songrong; sweet in taste, slightly warm. Treats the five taxations and seven injuries, tonifies the center, eliminates cold-heat pain in the genital tract, nourishes the five viscera, strengthens yang, benefits essence and qi, promotes fertility; for women's abdominal masses. Long-term use lightens the body. Grows in mountain valleys." This reflects early understanding of its main medicinal values. The origin is described as "mountain valleys," but specific original species and host plants were not yet identified.
During the Wei, Jin, and Northern and Southern Dynasties, Wu Pu's Materia Medica recorded: "Also called Rou Songrong. Shennong and the Yellow Emperor: salty; Lei Gong: sour; Li: slightly warm. It grows on shaded mountain slopes in Hedong. Length three to four inches, cluster-forming. Or in Dai Prefecture and Yanmen. Collected from February to August, dried in the shade." This is the first explicit note of its flavor-profile as "salty," with sourness and slight warmth, consistent with the tonic properties of saline-origin drugs in TCM. It also describes morphology and growth: root length 3–4 inches, cluster-forming, suggesting growth in arid, saline-alkali or valley-desert habitats. Combining flavor, root length, and cluster-forming traits, this likely corresponds closely to what was later termed C. salsa. The Supplemental Records of Famous Physicians also states: "Grows in Hexi and Dai Prefecture/Yanmen"; Collected Annotations on the Materia Medica adds: "Sweet in taste... grows in Hexi mountain valleys and in Dai Prefecture/Yanmen. Collected on the fifth day of the fifth lunar month, shade-dried. Dai Prefecture/Yanmen belongs to Bingzhou; where there are many horses, there it is said to grow from the essence of wild horses dripping onto the ground. When fresh it resembles flesh; cooked as mutton stew, it is excellent for replenishing debility; can also be eaten raw. Found in abundance in Henan. The best now comes from Longxi: flattened in shape, soft and moist, many flowers, and sweet in taste. Next is from the northern regions: short in shape and fewer flowers. Bādong and Jianping also have it, but inferior." This is the first appearance of "flattened, soft, moist, many flowers, sweet taste" for the Longxi area, matching desert-type traits, while "short in shape and fewer flowers" better matches C. salsa. The text emphasizes morphological differences and geoherbal quality ranking: "the best comes from Longxi; next from the northern regions." It also notes harvest and shade-drying methods and emphasizes Hexi and Dai/Yanmen as main production areas.
In the Tang, Su Jing's Newly Revised Materia Medica inherits prior records and notes: "This annotation discusses herbaceous Congrong; Tao [Hongjing] did not see the fleshy kind. What people now use is also the herbaceous Congrong with the flowers scraped off, used as a substitute. The original Classic has fleshy Congrong, and its efficacy is notably superior. In recent times, physicians occasionally use the true fleshy form." Before the Tang (Tao Hongjing's era), only the herbaceous "Congrong" was seen, and physicians mostly used it as a substitute for the fleshy Cistanche; the Divine Farmer's Classic recorded the true fleshy kind with stronger efficacy; by the Tang, the true fleshy Cistanche began to be used occasionally.
In the Five Dynasties, Rihuazi's Materia Medica recorded: "Cistanche... According to the Materia Medica, it is formed from the residual drippings of wild horse essence. After field inquiry, it was found beneath fallen trees, in well pits and earthen trenches-places not of horse copulation. Tao's account is mistaken. There is also 'flower Congrong,' that is the spring tender shoots, whose potency is relatively weak." While inheriting early records, Rihuazi corrected folk and ancient myths ("wild horse essence") through field observation, and first clearly distinguished "flower Congrong" (tender shoots, weaker) from the true fleshy Cistanche (mature, stronger). Shu Materia Medica states: "Produced in the sands of Fuxian, Suzhou (Gansu). The root is dug in March–April; cut out the central good parts, 3–4 inches, and string to shade-dry. It is best in the eighth month. The skin has scales like pinecone scales; the root exceeds a foot in length. As for the herbaceous Congrong, harvested mid-April, length 5–6 inches to over a foot, stem purple; after harvesting, press flat to dry in the sun. Yuanzhou, Qinzhou, and Lingzhou all have it." This further documents differences between fleshy and herbaceous forms in origin, harvest time, stem color, and processing, indicating that by the Five Dynasties, pharmacopoeial practice had made systematic observations and classifications of Congrong-type materials.
Song dynasty records systematically documented flavor, origin, morphology, and processing. The Kaibao Materia Medica follows earlier accounts, emphasizing: "The best now comes from Longxi-flattened, soft and moist, many flowers, sweet in taste. Next is from the northern regions-short in shape, fewer flowers. Bādong and Jianping also have it, but inferior." The Formulary of the Imperial Pharmacy notes major origins as "Hexi mountain valleys and Dai Prefecture/Yanmen." Both the Illustrated Materia Medica and the Extended Meanings of the Materia Medica recorded the morphological descriptor "skin with scales like pinecone scales." The Illustrated Materia Medica gave detailed accounts of morphology, harvest, and geoherbal regions: "Cistanche grows in the Hexi mountain valleys and Dai Prefecture/Yanmen; it is now found in many prefectures of Shaanxi, yet is not as good as that from within the Western Qiang border-its flesh is thick and strength firm. The old tale says it arises from the drippings of wild mares; now Westerners say it often grows among big trees and in earthen trenches and walls-places not of mares in heat-thus we know it has its own kind... Skin like pinecone scales. A slender, flat root beneath the shoot, longer than a foot. Harvest the root in the third month, digging the best central parts..." It also notes: "Harvest on the fifth day of the fifth lunar month. In May it may already be too old to use, hence many harvest in March... There is also an herbaceous kind very similar, but with short root, round stem, purple in color; people often scrape off the flowers and press it flat to substitute for the fleshy kind, but the efficacy is markedly inferior." Under the lower-grade "Orobanche" entry: "Grows on rocky slopes south of the mountains, like lotus root; dug when first sprouting, shade-dried; also called 'herb Congrong'..." Here, fleshy Cistanche from Longxi and Western Qiang is designated superior-flattened, rich in oils, with scaly skin and stronger efficacy-whereas the herbaceous form has short root, round purple stem, is often scraped and pressed to mimic the fleshy product, but is inferior.
Ming dynasty works, while inheriting Song foundations, provided more systematic accounts of origin, morphology, processing, and adulteration. Several texts indicate it was "born of horse essence," emphasizing the folk origin of the legend; Five Miscellanies records "skin like pine scales, soft and moist like flesh," further describing superior geoherbal morphology; The New Compilation of the Materia Medica emphasizes medicinal standards: "When using Cistanche, one must select those plump and with scales to be acceptable." Meanwhile, Yihu notes the true product is rarely obtained and that the market often sells pressed or salt-processed roots of other plants (e.g., Gentiana macrophylla) as counterfeits, with harmful effects if misused, highlighting the importance of authentic identification. Ben Cao Cheng Ya Ban Ji records detailed origin and morphological features: "Produced in the mountain valleys of Hexi and in Dai Prefecture/Yanmen; many prefectures in Shaanxi have it. It grows in clusters beneath fallen trees and along earthen trenches. In spring it sends up shoots, flesh-colored and reddish with scales. The best from Longxi is flattened, yellow-red, soft and moist, many flowers, sweet and plump; next is from the northern regions, short with fewer flowers. Bādong and Jianping also have it, but not good."
Qing dynasty literature continued prior understanding and added detail on material morphology and processing. Ben Cao Chong Yuan and Ben Cao Cong Xin state: "People now often use the herbaceous Congrong, whose potency is somewhat weaker"; Ben Cao Qiu Zhen and Ben Cao Bei Yao add descriptors of superior quality: "As large as an arm, weighing up to a catty or more, with pinecone-like scales is best." Donghua Lu records: "The two regions of Gansu are unlike Zhili and Shandong; like in Mongolia, even in years of famine, livestock can survive on dairy, and Cistanche is also produced locally," indicating Gansu's unique environment is suitable for Cistanche growth and its drought tolerance and desert adaptation. The Annals of Shaanxi note that Cistanche is widely produced in northern Shaanxi and is often used as food by local people. Ben Cao Fen Jing reiterates that the herbaceous form is weaker, reflecting awareness of adulterants. Illustrations and Explanations of Plants and Names state: "People often take the herbaceous Congrong to substitute for the fleshy one... all salted, scaly, flattened, blackish, and soft." Ben Cao Shu Gou Yuan records: "The superior Cistanche is described as soft red, beautiful and full, with oily luster coagulated-indeed flesh-like... The Longxi product is flattened, yellow and moist, many flowers and sweet; others are inferior." Geoherbal regions are identified as Longxi, Hexi, and prefectures of Shaanxi as superior.
2.2 Morphological Study
By comparing habitat, hosts, scale leaves, floral morphology, corolla color, and filaments of C. deserticola and C. salsa across China, we summarize as follows: C. deserticola plants are taller; leaves ovate or oblong-ovate; spikes with sparser upper leaves, lanceolate or narrowly lanceolate, glabrous; bracts linear-lanceolate or lanceolate, usually longer than the corolla; bracteoles ovate-lanceolate or lanceolate, nearly as long as the calyx; calyx campanulate, shallowly 5-lobed; corolla tubular-campanulate, with 5 lobes, nearly semicircular; corolla pale yellow (also occasionally pale purple), lobes pale yellow, pale purple, or with lilac margins; filaments at base with wrinkled long villous hairs; anthers with an abrupt apical tip, with wrinkled long villous hairs; nectary present at ovary base; stigma apex inflexed; capsule ovoid, with a persistent style at the apex. C. salsa plants are shorter; leaves ovate or oblong-ovate; spikes with ovate-lanceolate bracts; bracteoles nearly as long as the calyx; calyx campanulate, shallowly 5-lobed; corolla tubular-campanulate, tube white, shallowly 5-lobed, lobes semicircular and pale purple; anthers long-ovate, with a small apical tip, and the filament bases densely covered with wrinkled long villous hairs; ovary glabrous; capsule ovoid or elliptic.
In C. deserticola (Figure 3A), the flowers are overall fuller; the corolla is tubular-campanulate, the tube milky white with lobes purplish-brown; the calyx is broadly campanulate with nearly rounded lobes; there are four stamens with filaments bearing long villous hairs at the base; anthers with a short-pointed connective; multiple bracts (ovate-lanceolate) and bracteoles are visible, and the overall floral components (corolla, calyx, stamens, etc.) are more complete. In C. salsa (Figure 3B), the corolla is more open and spreading, the tube milky white with yellowish markings, the lobes are slightly lighter purplish-brown; calyx lobes are narrower (nearly lanceolate); bracts and bracteoles are fewer and more slender; filaments are relatively thinner; the overall floral size is slightly smaller than that of C. deserticola, with components arranged more loosely. Both species share the typical floral structure of parasitic Orobanchaceae, but can be distinguished by corolla texture, calyx morphology, and floral component details.
Comparing cross-sections of four Chinese Cistanche species (Figures 4–5), we found that both C. deserticola and C. salsa show a distinct radial structure in cross-section, with vascular bundles forming a ring of deep crenulate waves and a conspicuous pith. This structure may support taller stem growth and be related to efficient water and nutrient transport. Although C. sinensis also has a conspicuous pith, its cambial ring is simpler than in C. deserticola and C. salsa, suggesting slightly lower capacities for water transport and storage. C. tubulosa exhibits a star-dotted vascular bundle pattern, markedly different from the other three, indicating notable differences in host utilization strategies and ecological adaptation.
2.3 Molecular Systematics Based on Three Genomic Marker Sets
A phylogenetic tree constructed from shared plastid-encoded protein sequences of 32 Orobanchaceae species revealed three major clades: Orobanche, Cistanche, and Rehmannia. The tree also included Aphyllon, Conopholis, Castilleja, and Triaenophora (Figure 6A). Notably, five Cistanche species formed four subclades. Clade I comprises C. sinensis, with 64% support. Clade II comprises C. deserticola and C. salsa, which cluster together with 100% support. Clade III is C. tubulosa, which clusters with Clade II with 78% support. Clade IV is Cistanche phelypaea (L.) Cout., an introduced Cistanche species.
A phylogenetic tree based on shared mitochondrial-encoded proteins from 11 Lamiales species revealed six major clades: Orobanchaceae, Phrymaceae, Lamiaceae, Lentibulariaceae, Oleaceae, and Gesneriaceae. Notably, the three Cistanche species sampled formed two clades. Clade I comprises C. deserticola and C. salsa, which cluster together with 100% support. Clade II is C. tubulosa, which clusters with Clade I with 100% support (Figure 6B). In addition, Castilleja paramensis (also Orobanchaceae) is the closest relative to Cistanche with 100% support. Phylogenetic analyses of nuclear gene sequences likewise indicate a close relationship between C. deserticola and C. salsa (Figure 6C). A core and highly consistent finding is the exceptionally close phylogenetic relationship between these two species.
